I'm writing an extension that would ideally read files after they are downloaded (through the normal download process). Is that possible? I can get the filename through chrome.downloads operations, but I can't find ways of reading the actual bytes. I don't need to write or move them, just read.
I'm writing an extension that would ideally read files after they are downloaded (through the normal download process). Is that possible? I can get the filename through chrome.downloads operations, but I can't find ways of reading the actual bytes. I don't need to write or move them, just read.
Share Improve this question asked Mar 13, 2016 at 11:57 Giao LacavaGiao Lacava 1,82315 silver badges26 bronze badges 4- Ideally you should not be able to read any file from the filesystem. That would be a security breach. If the file is being downloaded/ stored in the temporary FileSystem used by HTML5 then it is possible. – user3613129 Commented Mar 13, 2016 at 12:09
- @user3613129 do you think I could somehow intercept a download to manage it with the fs api then? – Giao Lacava Commented Mar 13, 2016 at 12:30
- I don't think so. Again if your extension would be able to intercept the download of a user, that would also be a security breach. FS API has a different use case. It is used to write/ read to/ from the Local filesystem, where the local filesystem means a temporary protected path created by the browser for these operations. Please go through the FS API for more information. – user3613129 Commented Mar 13, 2016 at 12:36
- 1 @user3613129 Anything an extension can do can be seen as a "security breach". They have elevated privileges pared to the web code. – Xan Commented Mar 13, 2016 at 13:02
3 Answers
Reset to default 10I struggled with the exact same problem and finally found an easy workaround to read the content of downloaded files from a Chrome extension.
You just need to add a permission to file://*in your manifest file and once you have your file's path in the system (with chrome.downloads.search in the filename property), make a GET XMLHttpRequest to the url file://{filepath}.
No, the Downloads API does not provide a way to access the file contents.
The closest for interception would be WebRequest API, but at the moment that doesn't allow response access either.
The most you can do is to catch a request with webRequest, cancel it and try to XHR it yourself, storing the result in some temporary place like the mentioned FileSystem API. This would be a terrible UX though (the original request fails inexplicably) and may not work in all cases (as XHR can't set all the headers to fully replicate a request).
WhiteFangs' solution worked for me. HOWEVER, I should add the caveat that the XMLHttpRequest using the file protocol only worked for me when I initiated it from the background script. I got errors when I tried it from a content script or from the javascript environment of my browser action.